Face Generation

In this project, you'll use generative adversarial networks to generate new images of faces.

Get the Data

You'll be using two datasets in this project:

  • MNIST
  • CelebA

Since the celebA dataset is complex and you're doing GANs in a project for the first time, we want you to test your neural network on MNIST before CelebA. Running the GANs on MNIST will allow you to see how well your model trains sooner.

If you're using FloydHub, set data_dir to "/input" and use the FloydHub data ID "R5KrjnANiKVhLWAkpXhNBe".

In [1]:
data_dir = './data'

# FloydHub - Use with data ID "R5KrjnANiKVhLWAkpXhNBe"
#data_dir = '/input'


"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import helper

helper.download_extract('mnist', data_dir)
helper.download_extract('celeba', data_dir)
Found mnist Data
Found celeba Data

Explore the Data

MNIST

As you're aware, the MNIST dataset contains images of handwritten digits. You can view the first number of examples by changing show_n_images.

In [2]:
show_n_images = 25

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
%matplotlib inline
import os
from glob import glob
from matplotlib import pyplot

mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'mnist/*.jpg'))[:show_n_images], 28, 28, 'L')
pyplot.imshow(helper.images_square_grid(mnist_images, 'L'), cmap='gray')
Out[2]:
<matplotlib.image.AxesImage at 0x1096181d0>

CelebA

The CelebFaces Attributes Dataset (CelebA) dataset contains over 200,000 celebrity images with annotations. Since you're going to be generating faces, you won't need the annotations. You can view the first number of examples by changing show_n_images.

In [3]:
show_n_images = 25

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'img_align_celeba/*.jpg'))[:show_n_images], 28, 28, 'RGB')
pyplot.imshow(helper.images_square_grid(mnist_images, 'RGB'))
Out[3]:
<matplotlib.image.AxesImage at 0x1094dd898>
In [4]:
mnist_images[0].shape
Out[4]:
(28, 28, 3)
In [5]:
28*28*3
Out[5]:
2352

Preprocess the Data

Since the project's main focus is on building the GANs, we'll preprocess the data for you. The values of the MNIST and CelebA dataset will be in the range of -0.5 to 0.5 of 28x28 dimensional images. The CelebA images will be cropped to remove parts of the image that don't include a face, then resized down to 28x28.

The MNIST images are black and white images with a single color channel while the CelebA images have 3 color channels (RGB color channel).

Build the Neural Network

You'll build the components necessary to build a GANs by implementing the following functions below:

  • model_inputs
  • discriminator
  • generator
  • model_loss
  • model_opt
  • train

Check the Version of TensorFlow and Access to GPU

This will check to make sure you have the correct version of TensorFlow and access to a GPU

In [6]:
"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
from distutils.version import LooseVersion
import warnings
import tensorflow as tf

# Check TensorFlow Version
assert LooseVersion(tf.__version__) >= LooseVersion('1.0'), 'Please use TensorFlow version 1.0 or newer.  You are using {}'.format(tf.__version__)
print('TensorFlow Version: {}'.format(tf.__version__))

# Check for a GPU
if not tf.test.gpu_device_name():
    warnings.warn('No GPU found. Please use a GPU to train your neural network.')
else:
    print('Default GPU Device: {}'.format(tf.test.gpu_device_name()))
TensorFlow Version: 1.0.1
/Users/steffen/miniconda3/envs/tensorflow/lib/python3.5/site-packages/ipykernel_launcher.py:14: UserWarning: No GPU found. Please use a GPU to train your neural network.
  

Input

Implement the model_inputs function to create TF Placeholders for the Neural Network. It should create the following placeholders:

  • Real input images placeholder with rank 4 using image_width, image_height, and image_channels.
  • Z input placeholder with rank 2 using z_dim.
  • Learning rate placeholder with rank 0.

Return the placeholders in the following the tuple (tensor of real input images, tensor of z data)

In [7]:
import problem_unittests as tests

def model_inputs(image_width, image_height, image_channels, z_dim):
    """
    Create the model inputs
    :param image_width: The input image width
    :param image_height: The input image height
    :param image_channels: The number of image channels
    :param z_dim: The dimension of Z
    :return: Tuple of (tensor of real input images, tensor of z data, learning rate)
    """
    
    inputs_real = tf.placeholder(tf.float32, (None, image_width, image_height, image_channels), name='input_real')
    inputs_z = tf.placeholder(tf.float32, (None, z_dim), name='input_z')
    learning_rate = tf.placeholder(tf.float32, name='learn_rate')

    return inputs_real, inputs_z, learning_rate


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_inputs(model_inputs)
Tests Passed

Discriminator

Implement discriminator to create a discriminator neural network that discriminates on images. This function should be able to reuse the variabes in the neural network. Use tf.variable_scope with a scope name of "discriminator" to allow the variables to be reused. The function should return a tuple of (tensor output of the generator, tensor logits of the generator).

In [8]:
def discriminator(images, reuse=False, alpha=0.2):
    """
    Create the discriminator network
    :param image: Tensor of input image(s)
    :param reuse: Boolean if the weights should be reused
    :return: Tuple of (tensor output of the discriminator, tensor logits of the discriminator)
    """
    
    # also the computational time will be much higher, i decide to use the dcgan solution as a blue print
    # since the most important thing is to explain und understand the code and not simple copy+paste it, I
    # don't know, how I can show my learning progress here. :)
    
    # I adapt the code to the 28*28*3 dim of the images, use batch_normalization. also I  
    # implement the alpha variable for a leaky relu, so I will use "normal" relu
    
    with tf.variable_scope('discriminator', reuse=reuse):
        # Input layer is 28x28x3
        layer_1 = tf.layers.conv2d(images, 64, 5, strides=2, padding='same',
                                  kernel_initializer=tf.truncated_normal_initializer(stddev=0.1))
        relu_1 = tf.maximum(alpha * layer_1, layer_1)
        # 14x14x64
        #print(relu_1.get_shape())
        
        layer_2 = tf.layers.conv2d(relu_1, 128, 5, strides=2, padding='same',
                                  kernel_initializer=tf.truncated_normal_initializer(stddev=0.1))
        bn_2 = tf.layers.batch_normalization(layer_2, training=True)
        relu_2 = tf.maximum(alpha * bn_2, bn_2)
        # 7x7x128
        #print(relu_2.get_shape())

        layer_3 = tf.layers.conv2d(relu_2, 256, 5, strides=2, padding='same',
                                  kernel_initializer=tf.truncated_normal_initializer(stddev=0.1))
        bn_3 = tf.layers.batch_normalization(layer_3, training=True)
        relu_3 = tf.maximum(alpha * bn_3, bn_3)
        # 4x4x256
        # print(relu_3.get_shape())

        layer_4 = tf.layers.conv2d(relu_3, 512, 5, strides=2, padding='same',
                                  kernel_initializer=tf.truncated_normal_initializer(stddev=0.1))
        bn_4 = tf.layers.batch_normalization(layer_4, training=True)
        relu_4 = tf.maximum(alpha * bn_4, bn_4)
        # 2x2x512
        # print(relu_4.get_shape())
        
        # Flatten it
        flat = tf.reshape(relu_3, (-1, 2*2*512))
        
        logits = tf.layers.dense(flat,1)
        out = tf.sigmoid(logits)

    return out, logits


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_discriminator(discriminator, tf)
Tests Passed

Generator

Implement generator to generate an image using z. This function should be able to reuse the variabes in the neural network. Use tf.variable_scope with a scope name of "generator" to allow the variables to be reused. The function should return the generated 28 x 28 x out_channel_dim images.

In [9]:
def generator(z, out_channel_dim, is_train=True, alpha=0.2):
    """
    Create the generator network
    :param z: Input z
    :param out_channel_dim: The number of channels in the output image
    :param is_train: Boolean if generator is being used for training
    :return: The tensor output of the generator
    """
    # no pooling layer, since a pooling makes the representation approximatly invariant 
    # to small translations of the input (Goodfellow 2016)
    # this example is on mnist:
    # For object recognition it's helpful, since small changes like a rotated number, doesn't
    # effect the outcome. 
    # For GAN this invariance helps us to generate different outputs. with pooling the faces would 
    # look much more similar. This exact the opposite of our intention. 
    
    with tf.variable_scope('generator', reuse=not is_train):
        layer_1 = tf.layers.dense(z, 7*7*512)
        
        layer_1 = tf.reshape(layer_1, (-1, 7, 7, 512))
        # bn_1 = tf.layers.batch_normalization(layer_1, training=is_train)
        relu_1 = tf.maximum(alpha * layer_1, layer_1)
        
        layer_2 = tf.layers.conv2d_transpose(relu_1, 256, 5, strides=1, padding='same',
                                  kernel_initializer=tf.truncated_normal_initializer(stddev=0.1))
        bn_2 = tf.layers.batch_normalization(layer_2, training=is_train)
        relu_2 = tf.maximum(alpha * bn_2, bn_2)
        
        # print(relu_2.get_shape())
        #7x7x256 now
        
        layer_3 = tf.layers.conv2d_transpose(relu_2, 128, 5, strides=2, padding='same',
                                  kernel_initializer=tf.truncated_normal_initializer(stddev=0.1))
        bn_3 = tf.layers.batch_normalization(layer_3, training=is_train)
        relu_3 = tf.maximum(alpha * bn_3, bn_3)
        
        # print(relu_3.get_shape())
        # 14x14x128 now
        
        layer_4 = tf.layers.conv2d_transpose(relu_3, 64, 5, strides=1, padding='same',
                                  kernel_initializer=tf.truncated_normal_initializer(stddev=0.1))
        bn_4 = tf.layers.batch_normalization(layer_4, training=is_train)
        relu_4 = tf.maximum(alpha * bn_4, bn_4)
        
        # print(relu_4.get_shape())
        # 14x14x64 now
        
        # 28, 28, 32 goal
        
        # Output layer
        logits = tf.layers.conv2d_transpose(relu_4, out_channel_dim, 5, strides=2, padding='same',
                                  kernel_initializer=tf.truncated_normal_initializer(stddev=0.1))
        # 28x28x5 now
        # print(logits.get_shape())
        
        out = tf.tanh(logits)
    
    return out


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_generator(generator, tf)
Tests Passed

Loss

Implement model_loss to build the GANs for training and calculate the loss. The function should return a tuple of (discriminator loss, generator loss). Use the following functions you implemented:

  • discriminator(images, reuse=False)
  • generator(z, out_channel_dim, is_train=True)
In [10]:
def model_loss(input_real, input_z, out_channel_dim, alpha=0.2):
    """
    Get the loss for the discriminator and generator
    :param input_real: Images from the real dataset
    :param input_z: Z input
    :param out_channel_dim: The number of channels in the output image
    :return: A tuple of (discriminator loss, generator loss)
    """
    
    g_model = generator(input_z, out_channel_dim, alpha=alpha)
    d_model_real, d_logits_real = discriminator(input_real,reuse=False, alpha=alpha)
    d_model_fake, d_logits_fake = discriminator(g_model, reuse=True, alpha=alpha)

    d_loss_real = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_real, labels=tf.ones_like(d_model_real)))
    d_loss_fake = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_fake, labels=tf.zeros_like(d_model_fake)))
    g_loss = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_fake, labels=tf.ones_like(d_model_fake)))

    d_loss = d_loss_real + d_loss_fake
    
    return d_loss, g_loss


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_loss(model_loss)
Tests Passed

Optimization

Implement model_opt to create the optimization operations for the GANs. Use tf.trainable_variables to get all the trainable variables. Filter the variables with names that are in the discriminator and generator scope names. The function should return a tuple of (discriminator training operation, generator training operation).

In [11]:
def model_opt(d_loss, g_loss, learning_rate, beta1):
    """
    Get optimization operations
    :param d_loss: Discriminator loss Tensor
    :param g_loss: Generator loss Tensor
    :param learning_rate: Learning Rate Placeholder
    :param beta1: The exponential decay rate for the 1st moment in the optimizer
    :return: A tuple of (discriminator training operation, generator training operation)
    """
    t_vars = tf.trainable_variables()
    
    d_vars = [var for var in t_vars if var.name.startswith('discriminator')]
    g_vars = [var for var in t_vars if var.name.startswith('generator')]

    # Optimize
    with tf.control_dependencies(tf.get_collection(tf.GraphKeys.UPDATE_OPS)):

        d_train_opt = tf.train.AdamOptimizer(learning_rate, beta1=beta1).minimize(d_loss, var_list=d_vars)
        g_train_opt = tf.train.AdamOptimizer(learning_rate, beta1=beta1).minimize(g_loss, var_list=g_vars)
    
    return d_train_opt, g_train_opt


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_opt(model_opt, tf)
Tests Passed

Neural Network Training

Show Output

Use this function to show the current output of the generator during training. It will help you determine how well the GANs is training.

In [12]:
"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import numpy as np

def show_generator_output(sess, n_images, input_z, out_channel_dim, image_mode):
    """
    Show example output for the generator
    :param sess: TensorFlow session
    :param n_images: Number of Images to display
    :param input_z: Input Z Tensor
    :param out_channel_dim: The number of channels in the output image
    :param image_mode: The mode to use for images ("RGB" or "L")
    """
    cmap = None if image_mode == 'RGB' else 'gray'
    z_dim = input_z.get_shape().as_list()[-1]
    example_z = np.random.uniform(-1, 1, size=[n_images, z_dim])

    samples = sess.run(
        generator(input_z, out_channel_dim, False),
        feed_dict={input_z: example_z})

    images_grid = helper.images_square_grid(samples, image_mode)
    pyplot.imshow(images_grid, cmap=cmap)
    pyplot.show()

Train

Implement train to build and train the GANs. Use the following functions you implemented:

  • model_inputs(image_width, image_height, image_channels, z_dim)
  • model_loss(input_real, input_z, out_channel_dim)
  • model_opt(d_loss, g_loss, learning_rate, beta1)

Use the show_generator_output to show generator output while you train. Running show_generator_output for every batch will drastically increase training time and increase the size of the notebook. It's recommended to print the generator output every 100 batches.

In [13]:
def train(epoch_count, batch_size, z_dim, learning_rate, beta1, get_batches, data_shape, data_image_mode):
    """
    Train the GAN
    :param epoch_count: Number of epochs
    :param batch_size: Batch Size
    :param z_dim: Z dimension
    :param learning_rate: Learning Rate
    :param beta1: The exponential decay rate for the 1st moment in the optimizer
    :param get_batches: Function to get batches
    :param data_shape: Shape of the data
    :param data_image_mode: The image mode to use for images ("RGB" or "L")
    """
    
    _, image_width, image_height, image_channels = data_shape
    
    input_real, input_z, learn_rate = model_inputs(image_width, image_height, image_channels, z_dim) 
    
    d_loss, g_loss = model_loss(input_real, input_z, image_channels, alpha=0.2)
    
    d_train_opt, g_train_opt = model_opt(d_loss, g_loss, learning_rate, beta1)
    
    saver = tf.train.Saver()
    sample_z = np.random.uniform(-1, 1, size=(50, z_dim))
    
    samples, losses = [], []
    steps = 0
    
    with tf.Session() as sess:
        sess.run(tf.global_variables_initializer())
        for epoch_i in range(epoch_count):
            for batch_images in get_batches(batch_size):
                
                steps += 1
                
                batch_images = batch_images * 2
                batch_z = np.random.uniform(-1.0, 1.0, size=(batch_size, z_dim))
                
                _ = sess.run(d_train_opt, feed_dict={input_real: batch_images, input_z: batch_z, 
                                                     learn_rate: learning_rate})
                _ = sess.run(g_train_opt, feed_dict={input_real: batch_images, input_z: batch_z, 
                                                     learn_rate: learning_rate})

                if steps % 50 == 0:

                    train_loss_d = d_loss.eval({input_z: batch_z, input_real: batch_images})
                    train_loss_g = g_loss.eval({input_z: batch_z})

                    print("Epoch {}/{}...".format(epoch_i+1, epoch_count),
                          "Discriminator Loss: {:.4f}...".format(train_loss_d),
                          "Generator Loss: {:.4f}".format(train_loss_g))
                    # Save losses to view after training
                    losses.append((train_loss_d, train_loss_g))

                if steps % 50 == 0:
                    gen_samples = sess.run(
                                   generator(input_z, image_channels, is_train=False),
                                   feed_dict={input_z: sample_z})
                    samples.append(gen_samples)
                    _ = show_generator_output(sess, 16, input_z, image_channels, data_image_mode)

        saver.save(sess, './generator.ckpt')
    
    return losses, samples       

MNIST

Test your GANs architecture on MNIST. After 2 epochs, the GANs should be able to generate images that look like handwritten digits. Make sure the loss of the generator is lower than the loss of the discriminator or close to 0.

In [15]:
batch_size = 64
z_dim = 100
learning_rate = 0.0002
beta1 = 0.5


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 2

mnist_dataset = helper.Dataset('mnist', glob(os.path.join(data_dir, 'mnist/*.jpg')))
with tf.Graph().as_default():
    train(epochs, batch_size, z_dim, learning_rate, beta1, mnist_dataset.get_batches,
          mnist_dataset.shape, mnist_dataset.image_mode)
Epoch 1/2... Discriminator Loss: 1.6859... Generator Loss: 0.4312
Epoch 1/2... Discriminator Loss: 1.5596... Generator Loss: 0.5328
Epoch 1/2... Discriminator Loss: 1.5614... Generator Loss: 0.5189
Epoch 1/2... Discriminator Loss: 1.3005... Generator Loss: 0.6549
Epoch 1/2... Discriminator Loss: 1.1612... Generator Loss: 0.6934
Epoch 1/2... Discriminator Loss: 1.1541... Generator Loss: 0.7130
Epoch 1/2... Discriminator Loss: 0.9455... Generator Loss: 1.0779
Epoch 1/2... Discriminator Loss: 1.2260... Generator Loss: 0.5593
Epoch 1/2... Discriminator Loss: 0.9730... Generator Loss: 0.7700
Epoch 1/2... Discriminator Loss: 1.0900... Generator Loss: 0.6475
Epoch 1/2... Discriminator Loss: 1.2586... Generator Loss: 0.9502
Epoch 1/2... Discriminator Loss: 1.1636... Generator Loss: 0.7568
Epoch 1/2... Discriminator Loss: 1.3014... Generator Loss: 0.6598
Epoch 1/2... Discriminator Loss: 1.1180... Generator Loss: 0.7207
Epoch 1/2... Discriminator Loss: 1.1798... Generator Loss: 1.0439
Epoch 1/2... Discriminator Loss: 1.5948... Generator Loss: 0.3301
Epoch 1/2... Discriminator Loss: 1.0919... Generator Loss: 0.8006
Epoch 1/2... Discriminator Loss: 1.2269... Generator Loss: 0.7281
Epoch 2/2... Discriminator Loss: 0.9497... Generator Loss: 1.0014
Epoch 2/2... Discriminator Loss: 1.2152... Generator Loss: 0.7322
Epoch 2/2... Discriminator Loss: 1.1240... Generator Loss: 0.8949
Epoch 2/2... Discriminator Loss: 1.4041... Generator Loss: 0.4730
Epoch 2/2... Discriminator Loss: 1.2201... Generator Loss: 0.6648
Epoch 2/2... Discriminator Loss: 1.1087... Generator Loss: 0.7383
Epoch 2/2... Discriminator Loss: 1.4392... Generator Loss: 1.1974
Epoch 2/2... Discriminator Loss: 1.2705... Generator Loss: 0.6892
Epoch 2/2... Discriminator Loss: 1.2521... Generator Loss: 0.7888
Epoch 2/2... Discriminator Loss: 1.5163... Generator Loss: 0.3401
Epoch 2/2... Discriminator Loss: 1.2398... Generator Loss: 0.5975
Epoch 2/2... Discriminator Loss: 1.2338... Generator Loss: 0.7044
Epoch 2/2... Discriminator Loss: 1.1682... Generator Loss: 0.6829
Epoch 2/2... Discriminator Loss: 1.2128... Generator Loss: 1.1712
Epoch 2/2... Discriminator Loss: 1.2041... Generator Loss: 0.6972
Epoch 2/2... Discriminator Loss: 1.3110... Generator Loss: 0.6843
Epoch 2/2... Discriminator Loss: 1.0905... Generator Loss: 0.9628
Epoch 2/2... Discriminator Loss: 1.1543... Generator Loss: 0.8003
Epoch 2/2... Discriminator Loss: 1.3411... Generator Loss: 1.3927

CelebA

Run your GANs on CelebA. It will take around 20 minutes on the average GPU to run one epoch. You can run the whole epoch or stop when it starts to generate realistic faces.

In [16]:
batch_size = 64
z_dim = 100
learning_rate = 0.0002
beta1 = 0.5


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 1

celeba_dataset = helper.Dataset('celeba', glob(os.path.join(data_dir, 'img_align_celeba/*.jpg')))
with tf.Graph().as_default():
    train(epochs, batch_size, z_dim, learning_rate, beta1, celeba_dataset.get_batches,
          celeba_dataset.shape, celeba_dataset.image_mode)
Epoch 1/1... Discriminator Loss: 1.2109... Generator Loss: 0.6179
Epoch 1/1... Discriminator Loss: 0.5368... Generator Loss: 1.3694
Epoch 1/1... Discriminator Loss: 0.8117... Generator Loss: 1.0579
Epoch 1/1... Discriminator Loss: 0.9826... Generator Loss: 1.0455
Epoch 1/1... Discriminator Loss: 0.9326... Generator Loss: 1.0272
Epoch 1/1... Discriminator Loss: 1.0936... Generator Loss: 0.9526
Epoch 1/1... Discriminator Loss: 0.9626... Generator Loss: 1.1942
Epoch 1/1... Discriminator Loss: 1.2773... Generator Loss: 0.8324
Epoch 1/1... Discriminator Loss: 1.1025... Generator Loss: 1.0568
Epoch 1/1... Discriminator Loss: 0.8397... Generator Loss: 1.4204
Epoch 1/1... Discriminator Loss: 0.6971... Generator Loss: 1.1684
Epoch 1/1... Discriminator Loss: 1.0394... Generator Loss: 1.0139
Epoch 1/1... Discriminator Loss: 1.8541... Generator Loss: 0.2275
Epoch 1/1... Discriminator Loss: 0.2831... Generator Loss: 2.7377
Epoch 1/1... Discriminator Loss: 0.2954... Generator Loss: 3.2771
Epoch 1/1... Discriminator Loss: 0.2907... Generator Loss: 3.5766
Epoch 1/1... Discriminator Loss: 1.9151... Generator Loss: 0.2060
Epoch 1/1... Discriminator Loss: 0.6908... Generator Loss: 1.1048
Epoch 1/1... Discriminator Loss: 0.1034... Generator Loss: 3.3260
Epoch 1/1... Discriminator Loss: 0.2364... Generator Loss: 3.7355
Epoch 1/1... Discriminator Loss: 0.4446... Generator Loss: 2.2941
Epoch 1/1... Discriminator Loss: 0.2277... Generator Loss: 6.2529
Epoch 1/1... Discriminator Loss: 0.0763... Generator Loss: 5.6545
Epoch 1/1... Discriminator Loss: 0.0555... Generator Loss: 5.9243
Epoch 1/1... Discriminator Loss: 0.2257... Generator Loss: 2.6516
Epoch 1/1... Discriminator Loss: 0.1807... Generator Loss: 5.7466
Epoch 1/1... Discriminator Loss: 0.2275... Generator Loss: 3.4704
Epoch 1/1... Discriminator Loss: 0.2245... Generator Loss: 6.3934
Epoch 1/1... Discriminator Loss: 0.1144... Generator Loss: 3.5990
Epoch 1/1... Discriminator Loss: 0.8791... Generator Loss: 0.6901
Epoch 1/1... Discriminator Loss: 0.1754... Generator Loss: 3.0876
Epoch 1/1... Discriminator Loss: 0.7769... Generator Loss: 0.8637
Epoch 1/1... Discriminator Loss: 0.1508... Generator Loss: 3.0571
Epoch 1/1... Discriminator Loss: 0.1708... Generator Loss: 6.1587
Epoch 1/1... Discriminator Loss: 0.1490... Generator Loss: 5.3758
Epoch 1/1... Discriminator Loss: 0.3640... Generator Loss: 3.1153
Epoch 1/1... Discriminator Loss: 0.0710... Generator Loss: 5.2281
Epoch 1/1... Discriminator Loss: 0.2170... Generator Loss: 6.0592
Epoch 1/1... Discriminator Loss: 0.1925... Generator Loss: 3.3084
Epoch 1/1... Discriminator Loss: 0.8518... Generator Loss: 6.9405
Epoch 1/1... Discriminator Loss: 0.5150... Generator Loss: 2.5722
Epoch 1/1... Discriminator Loss: 0.2962... Generator Loss: 2.2651
Epoch 1/1... Discriminator Loss: 0.2497... Generator Loss: 2.2917
Epoch 1/1... Discriminator Loss: 0.1284... Generator Loss: 5.1908
Epoch 1/1... Discriminator Loss: 0.7368... Generator Loss: 5.6201
Epoch 1/1... Discriminator Loss: 0.6218... Generator Loss: 4.8107
Epoch 1/1... Discriminator Loss: 0.5943... Generator Loss: 2.7853
Epoch 1/1... Discriminator Loss: 0.6080... Generator Loss: 1.1100
Epoch 1/1... Discriminator Loss: 0.1662... Generator Loss: 2.7774
Epoch 1/1... Discriminator Loss: 0.1341... Generator Loss: 4.0235
Epoch 1/1... Discriminator Loss: 0.0510... Generator Loss: 5.7955
Epoch 1/1... Discriminator Loss: 0.2338... Generator Loss: 2.4613
Epoch 1/1... Discriminator Loss: 0.0854... Generator Loss: 7.7280
Epoch 1/1... Discriminator Loss: 0.2918... Generator Loss: 2.8131
Epoch 1/1... Discriminator Loss: 0.1279... Generator Loss: 3.4059
Epoch 1/1... Discriminator Loss: 0.1775... Generator Loss: 3.0856
Epoch 1/1... Discriminator Loss: 0.1012... Generator Loss: 3.4282
Epoch 1/1... Discriminator Loss: 0.1245... Generator Loss: 6.4501
Epoch 1/1... Discriminator Loss: 0.6090... Generator Loss: 1.3853
Epoch 1/1... Discriminator Loss: 0.2109... Generator Loss: 4.3018
Epoch 1/1... Discriminator Loss: 0.1368... Generator Loss: 5.2654
Epoch 1/1... Discriminator Loss: 0.5362... Generator Loss: 1.6965
Epoch 1/1... Discriminator Loss: 1.5940... Generator Loss: 0.3269

Submitting This Project

When submitting this project, make sure to run all the cells before saving the notebook. Save the notebook file as "dlnd_face_generation.ipynb" and save it as a HTML file under "File" -> "Download as". Include the "helper.py" and "problem_unittests.py" files in your submission.